During the last decade, the econometric study of investment behaviour has developed from the descriptive construction of empirical investment equations, as first performed by J. Tinbergen in his well-known 19- publication: 'A Method and its Application to Investment Activity', to the empirical testing of increasingly explicit theories of production behaviour. It was principally D. Jorgenson who gave a great push to this intensified interest in micro-economic investment, with his famous 1965-paper: 'Anticipation and Investment Behaviour', in which he formulated a pure neoclassical model of investment behaviour under the conditions of a (simple) homogeneous production technology and perfectly competi­ tive markets. But, although the scope of the familiar flexible accelerator model was considerably extended by the introduction of relative factor prices, the resulting investment relationship generally remained one of the most ill-estimated equations in econometric models. The very rigid assumptions of pure neoclassical models might have caused these bad results. More­ over, the required ex post measure for aggregate capital stock is very deficient for most economies. Hence, it should be interesting to formulate investment models subject to less rigid restrictions on production and market behaviour and, preferably, not containing any measure of aggre­ gate capital stock.

During the last decade, the econometric study of investment behaviour has developed from the descriptive construction of empirical investment equations, as first performed by J. Tinbergen in his well-known 19- publication: 'A Method and its Application to Investment Activity', to the empirical testing of increasingly explicit theories of production behaviour. It was principally D. Jorgenson who gave a great push to this intensified interest in micro-economic investment, with his famous 1965-paper: 'Anticipation and Investment Behaviour', in which he formulated a pure neoclassical model of investment behaviour under the conditions of a (simple) homogeneous production technology and perfectly competi­ tive markets. But, although the scope of the familiar flexible accelerator model was considerably extended by the introduction of relative factor prices, the resulting investment relationship generally remained one of the most ill-estimated equations in econometric models. The very rigid assumptions of pure neoclassical models might have caused these bad results. More­ over, the required ex post measure for aggregate capital stock is very deficient for most economies. Hence, it should be interesting to formulate investment models subject to less rigid restrictions on production and market behaviour and, preferably, not containing any measure of aggre­ gate capital stock.